cv

Data Engineer Intern

Prediko

Implemented ML pipeline for time series forecasting and developed dashboard for company and user activity monitoring.

• Implemented pipeline for time series forecasts using ARIMA, ARIMAX, SARIMAX and ensemble techniques with Darts library
• Developed full-stack dashboard (MongoDB, FastAPI, Docker, GCP) presented to investors for $5M funding round
• Used machine learning algorithms and Gaussian Processes for stock management optimization

Joint Ph.D.

Institute of Neuroinformatics, ETHZ and UZH

Electrical Engineering and Neuroscience

• Hardware-software co-design
• Dendritic computing
• Linear sequence models

Master of Science

ETH Zürich

Electrical Engineering and Information Technology

• Introduction to Machine Learning
• Advanced Machine Learning
• Neuromorphic Engineering I & II
• VLSI I, II & III
• Machine Learning on Microcontrollers
• Introduction to Neuroinformatics
• Deep Learning in Biological Networks

Bachelor

École Polytechnique Fédérale de Lausanne (EPFL)

Electrical and Electronic Engineering

• Analysis I, II, III, IV
• Physics I, II, III, IV
• Electrotechnics I & II
• Electronics I & II

Quantizing Small-Scale State-Space Models for Edge AI

ACM International Conference on Neuromorphic Systems (ICONS)

Demonstrated efficient low-bit state-space models through quantization-aware training with heterogeneous precision scheme reducing memory 6× without performance loss.

mGRADE: Minimal Recurrent Gating Meets Delay Convolutions for Lightweight Sequence Modeling

Under Review

Proposed a hybrid-memory architecture enabling efficient multi-scale temporal processing on edge devices with competitive results on long-range dependency tasks.

DenRAM: neuromorphic dendritic architecture with RRAM for efficient temporal processing with delays

Nature Communications

Developed hardware-aware training compatible with RRAM circuits for spiking neural networks achieving best performance for feedforward SNN models.